Retroviruses are viruses that contain a single strand of RNA as the genetic material rather than DNA. The single-stranded RNA genome of each of these viruses gives rise to a double-stranded DNA molecule after the virus infects a susceptible host. This DNA replica of the viral genome then introduces itself permanently into a chromosome of the successfully infected cell and replicates in that host chromosome.
The retroviruses discussed hereinafter and in the claims may be further defined as being replication-defective retroviruses. Thus, these viruses do not themselves contain a gene encoding the reverse transcriptase usually required to permit the viral RNA genome to be translated into a DNA that can be introduced into a chromosome fo the infected host. Rather, the retroviruses discussed hereinafter typically must be complimented in their infection by a so-called helper virus that is replication-competent. That second virus contains the gene that encodes the reverse transcriptase enzyme that incorporates the genomic materials from both viruses into the successfully infected host cells to transform those cells.
For ease in understanding, the replication-defective retroviruses will be discussed hereinafter and in the claims merely as retroviruses with the understanding that they are replication-defective and require the assistance of a helper virus for successful infection and transformation of host cells. This usage of the term retrovirus is known in the art and has been used in the art as such without further explanation.
Some members of the retrovirus family are highly oncogenic as judged by their ability to cause the formation of solid tumors within a short period of time after being inoculated into the host. These viruses can also cause "cancerous" changes in cells grown and cultured in the laboratory; such changes are called "transformations" and provide a reliable in vitro biological assay for oncogenic viruses. Several such viruses have been isolated from chickens, turkeys, mice, rats, cats and monkeys.
A single gene, the oncogene, located on the genome of these highly oncogenic viruses is responsible for the tumorgenic potential of the virus. In the case of several viruses, the protein products of their oncogenes, referred to herein as oncoproteins, have been immunologically identified by taking advantage of the fact that serum from an animal bearing a virus-induced tumor contains antibodies directed against those oncoproteins.
A rapidly growing body of evidence indicates that the oncogenes of retroviruses are closely related to and are derived from specific genetic loci in the normal cellular genetic information of all vertebrates. Molecular hybridization studies using specific nucleic acid probes done during the middle 1970's, followed by genetic cloning of viral oncogenes and their cellular relatives by recombindant DNA technology, have established the kinship between retroviral oncogenes (v-onc) and cellular oncogenes (c-onc) found in all normal vertebrate cells.
Molecular analysis of the nearly two dozen retroviruses thus far isolated has revealed more than a dozen different oncogenes, each distinguished by its nucleotide sequence, and each with a corresponding cellular oncogenic homolog. For example, the human EJ or T24 bladder carcinoma oncogene was identified as the homolog of the transforming gene of Harvey murine sarcoma virus (ras.sup.Ha) and also of the BALB sarcoma virus (bas) [Parada et al., Nature, 297, 474-478 (1982); Der et al., Proc. Natl. Acad. Sci. USA, 79, 3627-3634 (1982); and Santos et al., Nature, 298, 343-347 (1982)]. In addition, the oncogene of the human carcinoma cell line LX-1 was found to be homologous to the transforming gene of Kirsten strain of murine sarcoma virus (ras.sup.Ki) [Der et al., above]. Still further, the same v-onc for a c-onc designated fps of avian origin is represented at least twice among a limited number of avian retrovirus isolates; its mammalian cognate designated fes in feline species is found in two different strains of feline sarcoma viruses. Moreover, recent work has found a sequence homology between human platelet-derived growth factor (PDGF) and the oncoprotein encoded by the simian sarcoma oncogene, v-sis, and denominated p28.sub.sis [Antoniades et al., Science, 220, 963-965 (1983) and Devare et al., Proc. Natl. Acad. Sci. USA, 80, 731-735 (1983)].
The structural and immunological relatedness between the transforming sis gene product (p28.sup.sis) of simian sarcoma virus and platelet derived growth factor (PDGF) provides the most solid link between the transforming properties of oncogenes and the mitogenic action of growth factors. The sis gene is one of many oncogenes that have been transduced by retroviruses. These captured genes have been highly conserved through evolution, suggesting they serve important physiological functions. PDGF is a very potent mitogen for many connective tissue cell types in culture. It is stored in the alpha granules of platelets, and is released at sites of vascular damage. PDGF binding to specific cell surface receptors triggering a tyrosine-specific protein kinase activity. This event identifies a common mechanism used by a wide variety of growth factors and oncogenes.
Insulin, gastrin, epidermal growth factor (EGF) and transforming growth factors all bind to receptors that are associated with tryosine protein kinase activity. The oncogenes src, yes, fes, fps, ros, abl, fgr have tyrosine kinase activity while the oncogenes mos, raf, mht, and erb-B have sequence homology to kinase domain. Furthermore, sequence analysis of fragments of the EGF receptor demonstrate a very close homology with the predicted sequence of erb-B. Thus, the binding of a growth factor to a receptor with tyrosine kinase activity appears to be a common event in mitogenesis and transformation.
The precise molecular mechanisms of this interaction are not known. PDGF isolated from platelets contains two polypeptide chains that form disulfide bonded complexes that migrate on denaturing polyacrylamide gels between 20,000 to 35,000 daltons. Reduction destroys the biological activity of these complexes and produces proteins that migrate between 14,000 and 18,000 daltons. Sequence analysis of the material migrating at 18,000 daltons identifies two homologous but distinct sequences.
As discussed before, one of these sequences (PDGF-2) is highly homologus to the protein (p28.sup.sis) predicted by the nucleotide sequence of the simian sarcoma virus oncogene (sis). The homology begins at residue 67 and extends at least to residue 171. Recently, the isolation and sequencing of a human c-sis clone has extended this homology to the predicted carboxy-terminus. The open reading frame encoding the sequenced PDGF-2 region continues upstream, indicating PDGF isolated fom platelets is derived from a larger precursor, consistent with the 4.2 kb sis-related mRNA detected in various cell lines.
The protein encoded by the viral oncogene and having a corresponding, homologous protein within the host cell are both referred to herein as oncoproteins, although the cellular oncoprotein is typically present in small quantities in normal cells, and thus need not only be associated with neoplastic states. In addition, oncoproteins encoded by related oncogenes may have different molecular weights, e.g., the p85 and p108 oncoproteins encoded by v-fes.sup.ST and v-fes.sup.GA, respectively, and the 100-105k dalton protein of normal mink cells thought to be encoded by the c-fes gene. [Sen et al., Proc. Natl. Acad. Sci. USA, 80, 1246-1250 (1983).]The term oncoprotein is thus used generally herein for proteins whose genes and amino acid residue sequences are homologous, at least in part.
The oncoprotein is generally not present in the virus particle that infects the cell, but is only expressed after infection and transformation. The corresponding cellular oncoprotein is expressed at most minimally in normal cells and to a greater extent in neoplastic cells. Thus, the oncoprotein cannot typically be obtained from the virus. In addition, isolation of oncoproteins from cells is made difficult because of the small amount present, the complex mixture of proteins found in normal cells, and the relatively small amount of such proteins present even in transformed cells.
Oncoproteins encoded by v-onc and c-onc genes thus typically contain large sequences of amino acid residues that are homologous, but nevertheless are not usually identical. In addition, oncoproteins encoded by genes of different viral strains, each of which contains ostensibly the same oncogene, have been found to have slight variations in their amino acid residue sequences as exemplified above, and by the four published sequences of the ras gene which differ at the position of the twelfth amino acid residue. Thus, even when oncoproteins are in hand, it may be difficult to distinguish among them.
Immunologically induced receptor molecules such as monoclonal and polyclonal antibodies or the idiotype-containing portions of those antibodies are useful in purifying protein ligands to which they bind, as diagnostic reagents for assaying the presence and quantity of the protein ligands, as well as for distinguishing among homologous protein ligands.
The difficulties associated with obtaining quantities of oncoproteins typically militate against the preparation of receptors to those oncoproteins, although whole cell-induced monoclonal antibodies to v-fes and v-fps encoded oncoprotein have been reported by Veronese et al., J. Virol., 43, 896-904 (1982). In addition, even were whole proteins available for use as immunogens for inducing the production of such receptors, the use of large protein molecules as immunogens produces antisera containing polyclonal antibodies to the numerous epitopes of the large protein molecules.
Hybridoma and monoclonal antibody techniques utilizing whole proteins or large protein fragments as immunogens have been useful in narrowing the immunological response to such immunogens. However, such technology as heretofore practiced has been extremely time consuming and has provided only a relatively small number of hybridomas that secrete useful antibodies that recognize the immunogen. Moreover, even when successful, such techniques cannot be predictive of the chemical identity of epitope to which the receptor molecules are raised. Consequently, even after immunogen-recognizing receptors are produced, the obtaining of receptors to specific, chemically identified epitopic portions of the protein ligand has been a hit or miss operation that still further reduces the number of useful hybridomas that are ultimately produced.
Arnheiter et al., Nature, 294, 278-280 (1981) reported on the production of monoclonal antibodies that were raised to a polypeptide that contained 56 amino acid residues and corresponded in amino acid residue sequence to the carboxy-terminal portion of an intact interferon molecule. That 56-mer polypeptide thus corresponded to approximately one-third of the sequence of the intact molecule.
Arnheiter et al. reported on the production of eleven monoclonal antibodies. However, only one of those eleven monoclonal antibodies bound both to the polypeptide immunogen and also to the intact interferon molecule. In addition, that binding was not very strong as judged by the 3000-fold excess of intact interferon required to compete the antibody away from the synthetic polypeptide. None of the other monoclonal antibodies bound to the intact molecule.
In addition, the production of the hybridomas secreting those monoclonal antibodies required the spleens from three immunized mice. The low yield of the desired interferon-binding monoclonal antibodies, and the fact that three mouse spleens were needed for the preparation of those hybridoma cell lines indicates that those workers were relatively unsuccessful in their efforts.
Lerner et al. have been successful in obtaining protection of animals by the use of vaccines against pathogens by utilizing synthetic amino acid residue sequences of short to moderate length as immunogens. See Sutcliffe et al., Science, 219, 495-497 (1983).
However, it must be understood that until the present invention, successful preparation of hybridomas and their secreted monoclonal receptors differs from the successful preparation of a vaccine containing oligoclonal receptors. Thus, for a high yield monoclonal antibody preparation, it is necessary to stimulate B-cells to secrete large amounts of avid antibodies. On the other hand, for a synthetic vaccine, a wider spectrum of oligoclonal antibodies may be produced in smaller amounts and with lower avidities. In addition, protection of an animal against a pathogen typically requires both T-cell and B-cell activations so that a cellular response and a humoral response, respectively, can be induced in the animal.
A popular explanation for the success of synthetic polypeptide-containing vaccines in generating antibodies that recognize intact proteins and protect animal hosts involves a stochastic model in which the diversity of the immune response allows the observation of an infrequent event; i.e., the polypeptide adopting the conformation of its corresponding sequence in the native molecule. The concept that moderate-length polypeptides can frequently conform to native structures is contrary to theoretical and experimental studies. Rather, such polypeptides are thought to exist as an ensemble of a large number of transient conformational states that are in dynamic equilibrium. T-Cell activation by, and B-cell production of antibodies raised to, some of that conformational ensemble have been believed sufficient to provide protection upon vaccination.